Collection Of Accelerometry Data In A Large Prospective Cohort Study

PURPOSE: Self-reported measures of physical activity (PA) and sedentary time may be influenced by recall bias, social desirability bias, participant comprehension, and other sources of random and systematic error. Additionally, self-reported measures of physical activity largely focus on moderate-to-vigorous intensity physical activity (MVPA), but collection of other aspects of movement behavior, such as light intensity activity, and patterns and timing of physical activity and sedentary time via survey have largely been unsuccessful. As accelerometry-based measures are not subject to these biases, large-scale collection of these data will be crucial for our understanding of PA patterns and health. METHODS: Within the American Cancer Society’s Cancer Prevention Study-3 (CPS-3), a large nationwide prospective cohort study of ~300,000 U.S. adults, participants were invited to join the Accelerometry Substudy with a goal of collecting seven consecutive waking days of data on 20,000 participants. Consented participants wore the Actigraph GT3x + accelerometer on the hip aligning with the midline of the non-dominant thigh. Participants were also asked to record what time they woke up, put the device on, took the device off, and went to bed each day using a wear time log. Accelerometers were initialized at a sampling rate of 30 Hz without the low frequency extension and data were stored in one-second epochs for use in the Choi/Sojourn-3 axis algorithms. RESULTS: To date, accelerometry data have been processed for 17,081 of the ~21,000 participants (mean age 60 [SD ± 10] years, BMI 26.8 [SD ± 5.6] kg/m2, 71% women) who participated in the Substudy from 2019-2023. Average daily wear time was 14.75 hr./day and 97.8% of participants had 3+ wear days. On average, participants accumulated 36 min/day MPA, 11 min/day VPA, 291 min/day light PA, and 547 min/day sedentary time. CONCLUSIONS: Several large epidemiologic cohorts, including Women’s Health Study (n = 17,061) and Women’s Health Initiative (n = 6,382) have also collected hip-worn accelerometer data. Individually, accelerometry data from these cohorts will contribute to our understanding of complex PA behaviors and the relationships with mortality. These data can also be pooled to better understand relationships with cancer, cardiovascular, and other health outcomes.